Shared agency with agentic AI: A marketing module case

Generative AI is increasingly used in business schools for quick answers and essay drafting. However, literature suggests that prompt-based tools can encourage students to accept outputs rather than practise key skills such as critical thinking, decision-making, and taking responsibility for choices (Essien et al., 2024; Nguyen et al., 2024). This leaves a practical question for educators: how can we use AI to develop active, empowered learners rather than passive recipients?  

1. From AI to agentic AI in HE 

One promising direction is the rise of Agentic AI (AgAI): AI agents that can take initiative, adapt to context, and act proactively towards a goal, while still working with (and staying accountable to) instructors and students (Acharya et al., 2025). Thus, AgAI is a form of AI that can plan and carry out multi-step support for learning (from monitoring progress to suggesting next actions, and asking follow-up questions), not based exclusively on students prompts.  

Crucially, we see AgAI as supporting shared agency: (1) we set the goals, (2) we decide the boundaries, and (3) we can take or override any AI suggestion. The AI, in turn, can notice where we may be struggling, flag inconsistencies, and offer timely prompts that help us move forward successfully.  

What does this look like in practice? In our work, we outline a new way of collaboration between human (the learner or the instructor) and the AI (the algorithmic learning companion), namely the AgAI Enhanced Human-AI Collaboration Model with three basic areas of activity (Kremantzis et al., 2025): 

• Human sphere (students and instructors): We set the learning goals and the boundaries (for example, what counts as acceptable use of AI). We make the final decisions about what to do next.   

• AI sphere (the agent): The AI can monitor what is happening, analyse draft work, and suggest actions that fit the agreed goals and boundaries. This should be a pre-trained model (not trained by instructor/learners) that can be embedded in the learning material/tools (offered by the Schools). 

• Shared agency (working together): This is where the learning happens. The AI proposes an option or asks a question; we evaluate it, ask for clarification, revise our ideas, and move on. The aim is not “the AI gives the answer”, but “the AI helps us think”.  

This model enables highly adaptive, student-led learning. It keeps human agency central, personalizes support by sensing context in real time, and allows the AI to act proactively rather than reactively. Such features can enhance experiential learning, by supporting students as they work through realistic scenarios and providing personalised guidance at scale. 

2. How to embed AgAI in a business module: A marketing case  

We demonstrate this approach in a Marketing module case example in the International Foundation Year Programme (School of the Arts, Queen Mary University of London), illustrating how AgAI could be implemented in business education and the potential impact it could create. In this module, assessments have been redesigned to be more authentic and process-driven. Instead of a single end-of-term submission, students complete weekly, low-stakes tasks that build towards their final marketing research report, which is considered their first research-oriented project at university. Embedding AgAI requires deliberate planning across three stages among different stakeholders.  

Stage 1: Institutional and technical preparation (school-level support) 

The foundation for AgAI integration is an institutional environment that enables educators to work with AI in a governed, equitable, and pedagogically appropriate way. Key institutional actions include: 

• providing embedded AgAI tools within approved learning platforms, rather than relying on ad hoc or student-supplied systems; 

• constraining AI functionality to educationally appropriate roles, such as monitoring progression, comparing new work with prior decisions, and surfacing reflective questions; 

• explicitly excluding unauthorised uses (e.g., generating strategies, writing submissions, or making evaluative judgements). 
  

Stage 2: Pedagogical design (instructor-led) 

Once the institutional environment is in place, educators design how AgAI is used at the module level. At this stage, instructors: 

• identify the key decisions students are expected to make (e.g., business idea, target segment, positioning strategy); 

• define clear boundaries for AI use, distinguishing between what AI may support (e.g. coherence checking, questioning, prompting reflection) and what must remain human-led (e.g., strategic choice, final outputs, evaluation); 

• build visible evidence of decision-making into assessment design, such as decision logs, rationale statements, or reflections on AI use. 
  

Stage 3: Learner interaction design (shared agency in practice) 

Finally, the focus shifts to how students interact with AgAI during learning activities. AgAI support is structured around specific checkpoints, which means: 

• AgAI is activated during weekly low-stakes tasks, such as competitor analysis or draft marketing tools; 

• AI is prompted using two fixed inputs: earlier decision logs and the new artefact; 

• AI restates prior decisions, checks alignment, and poses a small number of clarification or reflection questions. 

Figure 1 illustrates how the proposed model operates in the marketing module. It visualises the interaction between three interconnected spheres across the weekly stages of students’ marketing research project. The figure highlights how responsibility is deliberately distributed: strategic ownership remains in the Human Sphere, analytical support sits within the AI Sphere, and learning occurs through dialogue and judgement in the Shared Agency space. 

Figure 1 Agentic AI (AgAI) Enhanced Human-AI collaboration in Marketing Module 

5. Impact on students learning 

AgAI has the potential to reshape students’ and academics’ experience in multiple way:  

• Student ownership of the learning process: Students remain at the centre of planning, developing, and refining their work, rather than responding to fragmented tutor inputs. 

• Skills development: Sustained ownership supports the development of skills in judging, evaluating, and critically aligning decisions with evidence and disciplinary concepts. 

• Purposeful progression: Timely AI prompts enable students to move forward effectively, progressing to next steps without waiting for lecturer rotation or revisiting misaligned work late. 

• Educators as strategic points: Educators withdraw from repetitive labour and focus on higher-level guidance, pedagogical judgement, and learning design. 

While illustrated here through a Marketing module, the approach has broader relevance across business education.  

1. Acharya, D. B., Kuppan, K., & Divya, B. (2025). Agentic AI: Autonomous Intelligence for Complex Goals - A Comprehensive Survey. IEEE Access, 13, 18912-18936. https://doi.org/10.1109/ACCESS.2025.3532853 

2. Essien, A., Bukoye, O. T., O’Dea, X., & Kremantzis, M. (2024). The influence of AI text generators on critical thinking skills in UK business schools. Studies in Higher Education, 49(5), 865-882. https://doi.org/10.1080/03075079.2024.2316881 

3. Kremantzis, M., Essien, A., Pantano, E., & Lythreatis, S. (2025). Uncovering the Generative AI (GenAI) to Agentic AI (AgAI) Shift for Business School Education. Journal of Global Information Management (JGIM), 33(1), 1-21. https://doi.org/10.4018/JGIM.389920 

4. Nguyen, A., Hong, Y., Dang, B., & Huang, X. (2024). Human-AI collaboration patterns in AI-assisted academic writing. Studies in Higher Education, 49(5), 847-864. https://doi.org/10.1080/03075079.2024.2323593